Study on global plant die-off faces questions

Kerry Sheridan AFP Yahoo News 26 Aug 11;

A study on plant productivity that said drought and global warming were killing off plants worldwide is now being questioned by scientists, according to research published Thursday.

In the study published in the journal Science last year, researchers Maosheng Zhao and Steven Running of the University of Montana used NASA satellite data to show that productivity declined slightly from 2000-2009.

Those findings contradicted previous studies from the 1980s and 1990s that showed warmer temperatures in some parts of the world were driving longer growing seasons and greater plant growth around the globe.

Having more plants on Earth would be good news because it would help offset greenhouse gas emissions by absorbing more carbon dioxide.

While Running noted at the time that the findings came as "a bit of a surprise," the study raised concerns about global food security, biofuels and our understanding of the carbon cycle.

The new questions about the study, published in Science on Thursday, are posed by scientists at Boston University in the United States and the Universities of Vicosa and Campinas in Brazil.

A press release distributed to reporters by Boston University said their study is "refuting earlier alarmist claims that drought has induced a decline in global plant productivity."

Statements included by the researchers describe Zhao and Running's model as "erratic," "poorly formulated," and showing no "trends that are statistically significant."

However, scientists who were not involved in either paper said this was an excellent example of the scientific process at work, and should not be cast otherwise.

"The Boston University press release -- using the term 'alarmist' -- speaks of a university trolling for media as distinct from a university seeking to communicate excellence," Andy Pitman, co-director of the Climate Change Research Centre at the University of New South Wales, said in an email to AFP.

"Those involved in this exchange of views are all well respected and excellent scientists. What is going on here is the scientific method. Zhao and Running publish a paper. Others attack it. Others defend it. Over time we determine who is right. Perfectly legitimate science."

Pitman, who has seen the study but was not involved in it, said the new analysis points to a smaller trend of plant loss but still shows declines over large swaths of territory in southeast Asia and China.

"This does not mean that there has been no decline, or that Zhao and Running's results were wrong, rather it highlights how strong research groups can reach different conclusions when using different assumptions," Pitman said.

"That opens up a rich vein of future research."

One of the key issues raised by critics was how the Zhao and Running study found a 0.34 percent reduction in the southern hemisphere's plant productivity, offset slightly by a 0.24 percent increase in the northern hemisphere, for a net decline of 0.1 percent over a 10-year period.

"This is the proverbial needle in a haystack," Simone Vieira, co-author and researcher at the State University of Campinas, Brazil, said in a statement.

"There is no model accurate enough to predict such minute changes over such short time intervals, even at hemispheric scales."

Lead author Arindam Samanta, a graduate of Boston University who is now at Atmospheric and Environmental Research Inc. in Lexington, Massachusetts, said the initial study's model was based on data from a decade when temperatures were on the rise.

"Their model has been tuned to predict lower productivity even for very small increases in temperature. Not surprisingly, their results were preordained," said Samanta.

According to NASA scientist Compton Tucker, who also reviewed the data, key questions to be resolved are whether the first study was accurate and whether its findings could be replicated over a longer period.

"It's just like studying the stock market for a few years versus 30 years," Tucker told AFP.

"Most people think you need a record of about 30 years of whatever data you are using in order to indicate a trend."

He said the publication of questions on the initial research should help advance knowledge in the area.

"This is science, where you take one step forward and two steps back," he said. "It's important for this to-and-fro to be able to play itself out."